Integrated Circuits and their evolution - Hybrid Circuits :
2. - OTHER LOGICAL FAMILIES
We talked about different technologies implemented to realize the logical functions. These are especially those that we meet most frequently.
There are others that must be known too, some are already part of ancient history.
a) The R.T.L.
(resistor - Transistor - Logic), resistance logic and transistors PNP in negative logic, is completely abandoned.
b) The R.C.T.L.
(Resistor - Capacitor - Transistor - Coupled - Logic),
resistor logic and transistors. The integrated capacitors were placed in parallel with the base resistors of the transistors to improve the signal transition speed.
c) The D.C.T.L.
(Direct - Coupled - Transistor - Logic),
direct coupled logic of transistors.
She is one of the first families, the two previous ones are derivatives.
These technologies were direct applications of traditional discrete-component assemblies. They have practically disappeared today.
However, from the bases of the latter (D.C.T.L), a new technology is realized :l'I.I.L.
d) L'I.I.L.
(or I²L
- Injection - Integrated - Logic)
or bipolar injection logic. This is a recent, low-speed technology with very low power consumption, which allows for great integration (we will discuss this problem in the chapter on integrated circuits)
This is a major asset in the future of a technology.
e) The T.D.T.L
(Tunnel - Diode - Transistor - logic),
tunnel diode logic and transistors.
The tunnel diode is certainly the fastest switching device known to date. This technology is very little used and is still being studied.
f) Unsaturated logic :
The
C.M.L. (Current
- Mode - Logic), current switching logic.
The
C.T.L.
(Complementary
- Transistor - Logic), logic with complementary transistors.
These different processes use bipolar transistors.
It is known on the one hand, the important in limiting the speed of operation is mainly due to the time taken by the transistors to leave the saturated state (storage-time).
Therefore, without affecting the structure of the transistors, it is sufficient to change the operating point thereof by ensuring that in the conductive state, it avoids the saturation zone.
Thus, the operators of this technology can operate at very high frequencies.
This is the logical mode, marketed, the fastest now.
Unfortunately, its noise immunity is very low (250 to 500 mV). Its dissemination to technicians is not very important.
According to the operators, the propagation time varies between 1.5 and 6 nanoseconds.
Consumption is about the same as that of TTL.
g) Threshold logic
Still called Threshold-Logic, it is practically not widespread, but it exists.
It has been designed for very specific applications in majority-decision digital circuits.
It is based on the principle of differential amplifiers, each input is assigned a different weight (or voltage threshold) and when half of the inputs plus one (that is, the majority) is in state 1 , the output changes state.
Several functions are thus realized, but are located away from the Boolean logic (one can say that it is logic with majority decision).
In addition, this technology is susceptible to noise, especially when the majority of inputs, minus one, are already in state 1.
The polarization resistors of the differential amplifiers must be of tight tolerances so that the thresholds are precise, which is an additional handicap to their industrial realization.
h) The devices C.T.D.
These load transfer elements (Charge - Transfert - Devices) are more specifically intended for registers, memories and delay lines, as well as for shooting devices.
They are based on the principle of injecting an electric charge into the semiconductor material.
This charge is then transferred using an electric field to the output.
This electric field is obtained by the application of a voltage system on gates (identical to those of the MOS transistors) distributed along the silicon bar.
Depending on the voltage distribution, the electric field propagates, causing the charge injected with it.
If the field remains stationary, the charge remains stored in the silicon bar.
The design of these devices is very simple and allows for great integration. One can, therefore, consider the development of this technology in the areas mentioned at the beginning of this paragraph.
Fig. 10. - The various technologies derived from silicon (those that are underlined are the most widespread).
Aviation was one of the first major consumers of electronic equipment. Imposed requirements forced manufacturers to reduce the weight and bulk of their devices.
The progress was especially spectacular with the advent of semiconductors.
Although the Cordwood technique is not part of the integrated circuits, it is good to mention it because, some achievements were of a rather exceptional density of components.
This method consists of placing two printed circuits parallel, one above the other, the coppered faces outwards. Passive components (usually resistors) serve as links and spacers to printed circuit boards.
In an assembly of this kind, well designed, inside the perimeter delimited by the edges of the printed circuits, it was possible to slip an additional component there, as small as it is.
Figure 11 shows, for information, a circuit of this kind.
This technique is the starting point of the race towards miniaturization.
3. 2. - THIN LAYERS
Passive components in fixtures occupy a very large volume.
On the other hand, it has long been known to apply conductive or insulating deposits on supports called substrates.
Research was therefore undertaken to reproducibly reproduce, on a substrate, the passive components required for the electronic circuits.
From this moment, we enter the field of microelectronics.
The substrates are generally glass or ceramic, on which the conductive, resistor or dielectric layers are successively deposited.
The substrate, mechanical support of any device, must have :
a coefficient of expansion chosen according to the use of the circuit.
a perfect surface condition.
a very high resistivity.
The choice of deposits is made according to their destination :
gold, copper, for connections
nichrome, tin oxide, for resistances
glass, silica, for dielectrics
silicon, for semiconductor layers
Different techniques are used to make these deposits. Those are :
silkscreen (made with a silk stencil). We will speak rather of thick layers with this process which does not allow to reach low thicknesses of deposits.
evaporation or vacuum spraying
gas deposition
electrolytic deposition
chemical deposition
In each of these techniques, the deposits are applied successively through an important process called : masks technique.
It consists of one or more masks with a photoresist followed by a chemical attack.
Figure 12-a traces the different phases of this process.
A first thin layer is deposited on the substrate, then a photosensitive product is applied to this layer.
It is then exposed to ultraviolet radiation through a mask or film.
After revelation and rinsing of this insolated layer, the underlying parts that must remain are protected by the photo-resistant product.
A chemical attack is then carried out to remove the parts of the deposit that must disappear.
At each deposit of a new layer, the same method is used by using a different mask.
Figure 12-b gives an overview of the realization of passive components in thin film technology (the scales are not respected).
The passive components are made in thin layers. If the locations and connections are provided on these circuits, active components such as transistors and diodes can be added.
The latter being presented in the form of micro-housings, they are placed and welded by micro-welds (see Figure 13).
By the same method, it is also possible to set up other passive components, such as capacitors whose value can not be obtained with the dielectric materials put in place using the deposits.
Small processors can also be implemented in the same way.
It only remains to protect everything mechanically, it is encapsulation. The capsule must include the output connections.
The size of hybrid circuits can range from a few millimeters to ten centimeters.
In general, flat boxes are used as soon as the size of the circuit reaches 1 to 2 cm. Below, the type transistors (type TO3, TO5 multi-output) are used.
The hybrid circuit technique is widely used in assemblies operating at very high frequency.
In general, it provides better performance and reliability compared to the traditional circuit (printed circuit wiring).
It is also used when the monolithic circuit (following paragraph) can not provide a solution, either because the dissipated power is too high, or because it involves integrating special components (transformer for example).
Integrated circuits and their evolution
The
C.M.L. (Current
- Mode - Logic), current switching logic.
Click here for the next lesson or in the summary provided for this purpose.
Top of page
Next Page 